This is an old revision of the document!

Weapon / Armor Information Reform

Name dcss:brainstorm:interface: combat info
Added by rob
Added on 2010-04-14 11:21

We're currently lacking good ways to aid the player in their choice of weapon.

Information currently provided explicitly:

  • weapon base speed
  • weapon base accuracy
  • weapon base damage
  • weapon brand
  • weapon skill

Other ingredients:

  • armour penalty (function of skills, armour and strength)
  • opponent ac/ev/resistances

Derived values:

  • effective weapon speed; git branch randomvar now shows this information (wizard mode only) — rob 2010-04-15 16:39
  • effective accuracy; wizard mode shows to-hit in '@'-description.
  • effective damage per hit
  • damage per normal-length turn
  • the above for a given opponent, e.g.
    • chance to hit MONSTER
    • average damage per hit against MONSTER
    • average damage per normal-length turn against MONSTER


  • For the random values, expected value (average), variance and extremes are interesting. E.g. effective weapon speed of 120% with a high variance / high maximum is worse than constant weapon speed of 120%.
  • Should think of Unarmed Combat; borsuk suggests to give UC a weapon-like description, perhaps access via “i-” (unarmed combat virtual “weapon” on inventory slot ”-”) 1344.
  • Most of this information is actually there for an observant player: By watching the hit/miss messages, you can get a feel for your accuracy in general and versus specific monsters. It's quite easy to tell how many turns it usually takes to kill a monster. It's more difficult to get information per absolute time, e.g. to see that a weapon of speed is about twice as good as the corresponding plain weapon. Conveying attack speed is probably the highest priority.
  • I think a major goal should be to allow an inexperienced player to answer this:
    • Is (Weapon A) currently better or worse than (Weapon B) for my character in speed, accuracy, and/or damage?
      • E.g. Someone finds a +5/+4 randart axe of electrocution and they want to determine if its a net gain to switch from the +1/+2 long sword they are skilled in
      • 1.5 handed weapons with or without a shield
      • Even comparing different weapons within a given weapon type given different brands, enchantments and base stats.
  • Suggestion: Show 3 bars (like spell power bars) accuracy, speed, damage. Its probably a good idea to use an exponential scale here like spell power already does. This way you can make basic comparisons between weapons unless they are too close (and then it might as well be up to the player to decide).
    • You can play with different resolutions here to provide a useful level of information without making it too spoiler-y.
    • This has the advantage that you don't have to mess around with providing either specific numbers nor coming up with a lot of text. I think its a simple but good solution to this problem. ryak
  • The information could get progressively more detailed as your weapon skill raises. When you reach a certain level, you get to know the average damage per turn. Later, you get the hit percentage. And for branded weapons, you could have damage against resistant or vulnerable monsters. And maybe also against heavy armoured or dodgy opponents, but that's probably going a little too far.
  • It doesn't have to be numbers, it could be bars or descriptions as already suggested. Or it could be vague descriptions at first, and precise numbers when you reach very high skill level.
  • Also, it would be nice if the system could be applied to spells, to answer questions like: “with my skills, should I keep using poison arrow against Pr monsters or should I switch to iron shot?” — galehar 2010-06-16 10:40

Messages / informational sentences

How do we convey the information? Numbers, messages, @-description, item description.

Some possible informational sentences:

  • Killing a plant takes 4 normal-length turns on average, and rarely more than 10.
  • Swinging this broad axe is a (little/lot) (slower/faster) than a normal-length turn. It could get a (bit/lot) faster with practice.
  • Hitting a killer bee is (very/slightly/ /not very/not) difficult.
  • add more, especially different ideas
This is something we should aim for. I am not sure about the usefulness of the first message (especially with the second half). Using killer bees as typical evasive monsters is good, not so sure about plants as typical health bags. (Note that monster EV is reliable, at least for bees, but monster health is not). However, I love the second and the third messages. We incorporated something like the third message in DCSS 0.2 (visible in @), but it was not very informative and was removed again. The only nitpick I have is the “normal-length turn”, but that can surely be circumvented. — dpeg 2010-04-14 23:46
I think the biggest problem with the 0.2 messages was that they mapped one value (tohit) to two variables (monster evasion, chance to hit monster of given evasion). — rob 2010-04-15 16:50

Related information

Idea graveyard

  • It might be interesting to allow the player to experiment with possible changes, e.g., how much would mastering the weapon skill or armour skill help? How about raising a stat? (Probably not possible to give this a good interface and to justify handing out the information in-game flavour-wise.)
    • I don't think we want any form of fsim in the game. See below for reasoning. I agree that we should get across the role of actual (not just base) speed better. In fact, it might be good to show fewer than the current three weapon numbers, and at the same time make the information more useful. — dpeg 2010-04-14 23:46
  • Make a polished fsim available to players. (Store a number of kits, allow choosing opponents, a nice in-game display of data like hit rate, number of hits required to kill monster, …).
  • Collect statistics on past fights. This would aggregate all melee attacks, and allow to filter these (how long ago, what opponent, what melee weapon, what skill level). The implementation would be quite generic, the interface should provide some sensible views of the data.
As you may have expected, I don't like either of these. The reason is that in whichever way we provide this information, players (also new ones) will feel they need it to win. Which is not the case. It is true that without such information, unspoiled players will always perform worse than spoiled ones. That's okay as long as the unspoiled players still can produce good enough decisions. (Which they can already now.) I absolutely support the idea of making it easier for players to come up with good decisions. There are some good ideas for that on this page. — dpeg 2010-04-14 23:46
Logged in as: Anonymous (VIEWER)
dcss/brainstorm/interface/combat_info.1324578773.txt.gz · Last modified: 2011-12-22 19:32 by XuaXua
Recent changes RSS feed Donate Powered by PHP Valid XHTML 1.0 Valid CSS Driven by DokuWiki